Hi!
I noticed that Proxmox lists CrateDB under success stories so I decided to give it a try.
Then I created a few very basic VM's for testing (maybe with suboptimal configuration but still):
10 CPU cores (1 x 10, no NUMA), 50 GB RAM (balloon = 0)
2 IDE disks where one is system disk and second is CrateDB data disk.
Options: Cache: Default (No Cache), Discard: ON, SSD Emulation: ON, Async IO: io_uring
File System: ext4 with mount option defaults,noatime,nobarrier (yes I forgot the discard option but could that be the issue?)
CentOS Linux 7.9.2009 (Core) kernel 3.10.0-1160.99.1.el7.x86_64, QEMU-Agent 10:2.12.0-3.el7
However after starting up the first node i got (really immediate, like literally the second crate service started):
Oct 19 11:24:05 crate: [2023-10-19T11:24:05,700][INFO ][o.e.n.Node ] [node1] started
Oct 19 11:24:05 kernel: JBD2: Spotted dirty metadata buffer (dev = sdb1, blocknr = 0). There's a risk of filesystem corruption in case of system crash.
...
Oct 19 11:24:07 kernel: EXT4-fs error: 1342 callbacks suppressed
Oct 19 11:24:07 kernel: EXT4-fs error (device sdb1): ext4_lookup:1447: inode #16515577: comm cratedb[node1][: deleted inode referenced: 17170609
Oct 19 11:28:10 kernel: EXT4-fs error (device sdb1): ext4_lookup:1447: inode #8388618: comm cratedb[node1][: deleted inode referenced:
After I fixed the filesystem with fsck.ext4 I tried to do some more testing with fio (randrw, read, write, ...) but could not find any more corruption.
pveversion: pve-manager/8.0.4/d258a813cfa6b390 (running kernel: 6.2.16-12-pve)
Any ideas what could possibly be the issue of this?
Many thanks in advance!
I noticed that Proxmox lists CrateDB under success stories so I decided to give it a try.
Then I created a few very basic VM's for testing (maybe with suboptimal configuration but still):
10 CPU cores (1 x 10, no NUMA), 50 GB RAM (balloon = 0)
2 IDE disks where one is system disk and second is CrateDB data disk.
Options: Cache: Default (No Cache), Discard: ON, SSD Emulation: ON, Async IO: io_uring
File System: ext4 with mount option defaults,noatime,nobarrier (yes I forgot the discard option but could that be the issue?)
CentOS Linux 7.9.2009 (Core) kernel 3.10.0-1160.99.1.el7.x86_64, QEMU-Agent 10:2.12.0-3.el7
However after starting up the first node i got (really immediate, like literally the second crate service started):
Oct 19 11:24:05 crate: [2023-10-19T11:24:05,700][INFO ][o.e.n.Node ] [node1] started
Oct 19 11:24:05 kernel: JBD2: Spotted dirty metadata buffer (dev = sdb1, blocknr = 0). There's a risk of filesystem corruption in case of system crash.
...
Oct 19 11:24:07 kernel: EXT4-fs error: 1342 callbacks suppressed
Oct 19 11:24:07 kernel: EXT4-fs error (device sdb1): ext4_lookup:1447: inode #16515577: comm cratedb[node1][: deleted inode referenced: 17170609
Oct 19 11:28:10 kernel: EXT4-fs error (device sdb1): ext4_lookup:1447: inode #8388618: comm cratedb[node1][: deleted inode referenced:
After I fixed the filesystem with fsck.ext4 I tried to do some more testing with fio (randrw, read, write, ...) but could not find any more corruption.
pveversion: pve-manager/8.0.4/d258a813cfa6b390 (running kernel: 6.2.16-12-pve)
Any ideas what could possibly be the issue of this?
Many thanks in advance!
